Plasma circular RNA panel to diagnose hepatitis B virus-related hepatocellular carcinoma: A large-scale, multicenter study

To explore whether plasma circular RNAs (circRNAs) can diagnose hepatitis B virus (HBV)-related hepatocellular carcinoma (HCC), microarray and qPCR were used to identify plasma circRNAs that were increased in HBV-related HCC patients compared to controls (including healthy controls, chronic hepatitis B and HBV-related liver cirrhosis). A logistic regression model was constructed using a training set (n = 313) and then validated using another two independent sets (n = 306 and 526, respectively). Area under the receiver operating characteristic curve (AUC) was used to evaluate diagnostic accuracy. We identified a plasma circRNA panel (CircPanel) containing three circRNAs (hsa_circ_0000976, hsa_circ_0007750 and hsa_circ_0139897) that could detect HCC. CircPanel showed a higher accuracy than AFP (alpha-fetoprotein) to distinguish individuals with HCC from controls in all three sets (AUC, 0.863 [95% confidence interval, CI: 0.819-0.907] vs. 0.790 [0.738-0.842], p = 0.036 in training set; 0.843 [0.796-0.890] vs. 0.747 [0.691-0.804], p = 0.011 in validation set 1 and 0.864 [0.830-0.898] vs. 0.769 [0.728-0.810], p < 0.001 in validation set 2). CircPanel also performed well in detecting Small-HCC (solitary, <= 3 cm), AFP-negative HCC and AFP-negative Small-HCC.